Pelion connects cloud-based IoT facial recognition security system, Facewatch

With over £1.5bn of retail goods disappearing from shelves due to theft every year and violence against staff increasing relentlessly, retailers are turning to smart deterrent solutions to prevent crime. Facewatch, the UK-based leaders in UKGDPR compliant IoT facial recognition, are enabling successful deployment to many of the leading brands in retail with their crime prevention service.

Retail crime continues to rise year-on-year, with police resources simply incapable of dealing effectively with extensive theft, threatening behaviour, and violence against staff. By introducing a Facewatch system in their stores, users can add Subjects of Interest who are reasonably suspected of having committed an offence using Facewatch’s high quality image detections and receive real-time alerts when those Subjects return to the store, enabling them to deal with the individual as appropriate.

Facewatch works by installing a dedicated CCTV camera at store entrances, connected to an edge device which detects all faces and transmits their images securely via a mobile 4G router using Pelion’s resilient IoT connectivity to their secure cloud. Their IoT facial regonition solution provides the ultimate in crime deterrence for their customers across the UK.

“Thieves don’t want to be caught; they don’t want to be recognised. In deployments that are rolled out nationally, our customers are seeing between 25% and 50% reduction in crime over the course of 12 months.”

Facewatch provides a holistic solution based on the business needs of their customer, handling every facet of the service on their behalf. This includes easy installation and setup, testing of IoT connectivity, fine-tuning of cameras, management of the facial recognition software and compliance with Data Protection regulations as data controllers. At the heart of their business, Facewatch is a data company, expertly transmitting, analysing, and processing the data being collected from their customer sites. The amount of data being transmitted is significant, with encrypted images constantly sent to the cloud, incidents being reported, and alerts sent to the Facewatch users.

“The challenge is that there’s a lot of data. Every time someone walks in, it’s not just one face or one image, but multiple frames per second from the video being processed. We need to get that to our cloud instantly and securely, where we process it within a fraction of a second to ensure full data protection compliance and complete privacy protection.”

As a hardware agnostic solution, the Facewatch platform and algorithms can run using any camera and router, connected with any SIM. They are unique in using two independent algorithms, combined with human experts to ensure their alerts are to the very highest standard of accuracy and integrity. They also provide robust post-sale service for their customers, working closely to keep them informed, sending them real-time management reports and helping them use the system as effectively as possible, to maximise their ROI.

The challenge

For the Facewatch system to work successfully, it needs extremely reliable connectivity, with solid coverage and low latency. The advantages of Facewatch's IoT facial recognition include its accuracy, speed, and data sovereignty. As the data produced never touches the customers’ networks, it’s completely isolated in terms of connectivity, and transmitted, processed, and stored independently.

The typical network infrastructure in retail uses copper broadband which is approximately 6x slower than Facewatch’s cellular-based solution for uploading data. For their customers to be satisfied, Facewatch must enable them to act quickly, with real-time alerts. And as they offer their solution as a subscription-based offering, it is absolutely essential for Facewatch to avoid slow service and downtime.

“When a subject walks into the store, there needs to be an alert in a couple of seconds. There’s no point in receiving an alert to potential behaviour when someone’s been in a store and already stolen goods; to be effective as a deterrent, they must be approached promptly.”

As a relatively new, rapidly expanding business, Facewatch were seeing significant growth and needed a partner who could support and accommodate them. Whereas large wholesale SIM providers work with companies based on volume, Facewatch felt that Pelion truly saw the opportunity in the business and industry and were invested in their success.

“We’d have customers that would get in contact with us, that had tested a couple deployments and realised quickly how fantastic the platform was and ask us if we can fit another 120 sites. We’d go to our provider, and they’d tell us that it would be a challenge to get the right SIMs in that quantity. It just wasn’t working.”

The solution

Pelion provides:

• Ultra-reliable, low latency connectivity over 4G, with excellent coverage regardless of location

• Competitive pricing and helpful alerts to avoid bill overages

• Excellent post-sales service and rapid access to support whenever needed

Facewatch needed a partner who could offer their sought-after combination of well-priced, robust connectivity and active support. They looked to Pelion as that partner, to properly integrate with their business and enable their growth.

“What we’re interested in is how fast can these guys help you when you need it? How effectively can they help you scale; how quickly can they provide support? That was where Pelion really impressed us.”

When there are any technical issues on-site, Facewatch have installers contacting them from across the UK and if they run into problems, they need support ASAP. Working with Pelion brings Facewatch the level of support, engineering, and troubleshooting that they’d been searching for. Once installation is complete, the finetuning of settings on the camera can be done remotely by a specialist, thanks to remote connectivity provided by the Pelion SIM in the Facewatch router, saving on time and resource.

Processing IoT facial recognition means high data usage

Avoiding bill shock was critical for Facewatch. For any business dealing in large volumes of data, variation of usage can be make or break for their profit margins. Pelion has worked closely with Facewatch in order to streamline as much of the data transmission as possible and offered competitive rates for their high usage. They’ve also enabled smart alerts which will inform them anytime a SIM is coming up to its data limit. Pelion supplies a detailed breakdown of costs in one unified platform, so no matter how much Facewatch scales up, they’ll be able to easily monitor usage and control every SIM to avoid overages.

Future plans

Edge intelligence is an exciting frontier for Facewatch, as they’re currently working with their partners to build the algorithms into the cameras, in order to speed up data processing. And the logical next step is to build the routing capability and the SIM card into the product as well, to supply one device that does everything they need.

Facewatch’s current plan is to significantly accelerate growth in the UK. They’re in discussions with large retailers across the country, with some big customers already using Facewatch in hundreds of locations.

Unsurprisingly, Facewatch’s adherence to GDPR compliance and their obsession with privacy protection for shoppers is one of the main attractions in an industry fraught with potential data privacy issues.

Nick Fisher explains, “In order to be UKGDPR compliant we need to comply with 3 main principles. First, you must demonstrate the necessity, that there is a true need for the data to be processed. Unfortunately for the retail industry, as violence against staff is rife and the cost of retail crime is around £1.5bn, there is a clear need. Facewatch works directly with their customers to ensure they meet the threshold to prove there is a need for their solution and that their platform is being used to satisfy that need.

Secondly, processing and storing the data is only permissible for the prevention of crime. And the data can only be shared proportionately in relation to the nature of the crime. So, if the individual was caught stealing bacon from a local shop, they shouldn’t end up on a national watchlist, but if they’re part of a criminal gang going around the country stealing high value goods, their data could justifiably be shared more widely.

Third, it must be in the 'substantial public interest', so private companies can’t simply install their own systems and hold all the data privately on their servers, as it doesn’t serve the substantial public interest. They need a data aggregator such as Facewatch to store the data securely on their behalf, process it and share it. If multiple shops are contributing to the data held by Facewatch, they all benefit from a larger database, which is a win for all the retailers. Facewatch takes all the responsibility for the storing and processing of data from their customers, reassuring them that it’ll be done in line with regulations and will be commercially confidential.”

Another exciting development for Facewatch's IoT facial recognition is the introduction of 5G cellular connectivity by Pelion, enabling better latency and equipping them with all the right technology for their continued scale out and growth.

“Our business is evolving; edge technology is rapidly becoming smarter. We just needed a partner that was prepared to support us in our journey.”
- Nick Fisher